Eur. Phys. J. Appl. Phys. 53, 10501 (2011) DOI: 10.1051/epjap/2010100298
THE EUROPEAN PHYSICAL JOURNAL APPLIED PHYSICS
Regular Article
Experimental study of the electrical properties of copper nitride thin films prepared by dc magnetron sputtering D. Dorraniana and G. Mosayebian Plasma Physics Research Center, Science and Research Branch, Islamic Azad University, Tehran, Iran Received: 31 July 2010 / Accepted: 27 September 2010 c EDP Sciences Published online: 23 December 2010 – Abstract. In this work the main effective parameters on the electrical resistivity of copper nitride thin films are investigated. Copper nitride thin films were successfully deposited on glass substrates by reactive dc magnetron sputtering at room temperature but different sputtering time. Working gas was a mixture of argon and nitrogen with equal amounts. The effect of deposition time on the structural, optical and electrical properties of deposited films was investigated. X-ray diffraction measurements show different lattice orientation in the structure of deposited films. By increasing the time of sputtering an orientation change from (100) to (111) can be observed in the films. Film morphology of samples is not changed with the sputtering time. The optical transmittance of deposited films decreases with increasing the deposition time. Results confirm that when the amount of nitrogen in working gas is 50%, we have more (100) planes in the structure of the deposited films, leads to higher resistivity of the films.
1 Introduction Copper nitride has a cubic anti Re-O3 type structure. In this structure, nitrogen atoms are positioned at the corners of the cell and copper atoms are positioned at the centre of the cube edge [1,2]. Copper nitrides have attracted considerable attention as a new material for optical storage devices and semiconductor integrate circuit, based on its unique properties, such as the rather low thermal decomposition temperature, semiconductor properties and excellent optical qualities [3,4]. The structure of the deposited copper nitride thin films is relatively stable; there was no change in condition of 95% humidity at a temperature of 60 ◦ C for 15 month. At very low temperature (about 300–470 ◦ C), copper nitride decomposes into Cu and N2 and becomes a conductor from a semiconductor. So it can be used as the write-once optical recording material upon heating. A number of nonequilibrium techniques, such as RF reactive sputtering, dc magnetic sputtering, ion assisted vapor deposition, reactive pulsed laser deposition, and other methods, are currently available for the preparation of copper nitride films [5–8]. However, the physical properties of the copper nitride films reported in literature are inconsistent. For example, the reported lattice constant of cubic Cu3 N ranges from 0.3815 to 0.3885 nm [5,9] or the optical band gap varies from 0.25 to 1.9 eV [10,11]. Electrical resistivity is a very important parameter of a semiconductor. A semiconductor with smaller electrical resistivity is closer to conductors with smaller bandgap a
e-mail:
[email protected]
energy, smaller transmittance with larger imaginary part of refractive index. Electrical resistivity of semiconductors of nitrides or oxides of transient metals can be influenced by the contents of nitrogen or oxygen in their lattice structure as well as its preferred orientation. In one of our previous experiments [12], unexpectedly the magnitude of the electrical resistivity of Cu3 N sample prepared at 50% nitrogen in the working gas was maximum between other ones deposited with other contents of nitrogen in the working gas. To found out more about this behavior, the influence of sputtering time on the structural and optical properties of Cu3 N films is investigated in this work, when the contents of argon and nitrogen in working gas are equal. Samples were deposited on glass substrate by reactive dc magnetron sputtering of Cu target at room temperature. Results confirm that the highest observed resistivity of Cu3 N thin films can be achieved in this experimental situation [2,5,7,12].
2 Experimental details Copper nitride thin films were deposited on BK7 glass substrate, by using a cylindrical direct current reactive magnetron sputtering system of a 99.99% pure copper. Sputtering reactor was consisted of two coaxial cylinders with 195 mm height. The diameter of the internal cylinder (cuprous cathode) was 30 mm, while glass substrates were placed on the outer cylinder with 60 mm diameter. Working gas was a constant mixture of 99.999% pure nitrogen and 99.999% pure argon with equal amounts at room temperature. The BK7 glass substrates were
Article published by EDP Sciences
The European Physical Journal Applied Physics
Fig. 2. Deposition rate of sputtering.
Fig. 1. X-ray diffraction spectrums of the deposited films.
11.34 n
cleaned by ultrasonic waves in alcohol and acetone for 10 min, and finally dried by blowing air. The sputtering system was capable of creating an ultimate vacuum of 2 × 10−4 mbar with rotary and diffusion pumps combination. During the sputtering experiment pressure was maintained at 2 × 10−2 mbar. The distance between the substrate and the target was 30 mm. The discharge current was 200 mA at the voltage difference of about 700 V. Uniform magnetic field of 400 Gauss was generated by a solenoid parallel to the axis of cylindrical chamber. Some features of sputtered films are presented in Table 1. After deposition, X-ray diffraction (XRD) was performed by STOE-XRD diffractometer using Cu-Kα line (l = 0.15406 nm), for phase identification and qualitative texture characterization. Change in transmittance and absorbance spectrums of the samples was measured by a Varian Cary-500 spectrophotometer at room temperature. A Dektak 3 profilometer was employed to measure the thickness of the growth films. Atomic force microscopy (AFM) micrographs were taken using Auto probe CP from park scientific Instrument and the electrical resistivity was measured by a two point probe system with Keithley 236 voltage source.
3 Results and discussion 3.1 Structure and morphology of films X-ray patterns of Cu3 N thin films are presented in Figure 1. X-ray photons diffracted mainly from (100) planes of Cu3 N at 2θ = 23.34◦, (111) planes at 2θ = 41.03◦ and (200) planes at 2θ = 47.74◦, confirm the polycrystalline structure of generated films. Result are similar with the results of our previous experiment reported in reference [12]. As is clear by increasing the deposition time from 9 to 21 min from sample 1 to 5, the contribution of (111) planes in the lattice structure is decreased while we have more intense peaks corresponds to (100) and (200) planes. (100) and (200) planes are N-rich planes while (111) planes are Cu-rich planes. By increasing the deposition time from 9 min to 15 min the intensity of peak corresponds to (111) Cu-rich planes is remained unchanged,
200nm 0.00 n
Fig. 3. (Color online) 2D AFM micrograph of the sample 2.
in other words number of sputtered Cu atoms does not change significantly. But after 15 min we have more N atoms in the lattice structure and preferred orientation of thin films changes to (100) and (200) planes which are N-rich planes. In fact increasing the time of deposition leads to more combination of nitrogen and copper atoms in the reactor rather than deposition of more Cu atoms. The approximately constant deposition rate of samples is presented in Figure 2 confirm that the target was not nitride in the time interval of the experiment. By increasing the deposition time we have more intense peaks of diffracted X-ray photons, show that the degree of crystallinity of films is increased. It can be due to the higher temperature of substrate at longer deposition times. For deposition time more than 21 min, the deposition rate decreases noticeably, confirms that at the applied pressure of this experiment with 50% nitrogen contents nitridation of target is very high. Using known Scherrer’s formula, the size of Cu3 N grains on the glass substrate are found to be between 36 to 48 nm for (100) planes, without any uniform variation, as is presented in Table 1, but this magnitudes are not confirmed by AFM pictures. 2D AFM micrograph of deposited film of sample 2 with the largest grains (according to AFM picture) is presented in Figure 3. As can be seen the size of grains vary from about 100 to 200 nm. 3d AFM micrographs of samples are very similar which are shown in Figure 4. We have nodule like structure films with similar nodule size, and large voided boundaries between
10501-p2
D. Dorranian and G. Mosayebian: Electrical properties of copper nitride thin films
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Fig. 4. (Color online) 3D AFM images of Cu3 N films. Table 1. Time film thickness and band gap energy of samples for deposited films. Sample Time (min) Film thickness (nm) Band gap energy (eV) Resistance (Ω) Grain size (100) planes (nm)
1 9 360 1.91 2 × 109 37.57
nodules. Usually for good crystalline films, a conical structure appears on the film surface. In this case the surface structure of the films confirms the presence of large amount of amorphous Cu3 N in the samples. In our previous work on Cu3 N also, when the content of nitrogen in working gas was 50%, in contrast with other samples, a nodule like structure was observed on the surface of film. The reason can be a topic of future work.
3.2 Electrical resistance of films The resistance of films was measured with a two point probes system using a Keithley 236 voltage source and an ampere meter in μA range. Applied voltage was varied from –100 to 100 V. The distance between contact needles was 1 mm. The diagram of applied voltage versus current for sample 5 is shown in Figure 5 and the electrical resistance of films is presented in Figure 6. Applied voltage
2 12 450 1.85 1 × 109 47.18
3 15 610 1.81 1 × 108 36.46
4 18 630 1.90 2 × 109 38.95
5 21 770 1.76 7 × 106 40.11
as a function of measured current for all samples is linear and resistance of deposited films is the slop of this line. By this method the resistance of films was measured between 7 × 106 to 2 × 109 Ω. As was seen for the case of deposited film with 50% nitrogen contents in the working gas [12] the electrical resistivity of films is higher than the reported magnitudes for Cu3 N. Several reasons can be responsible for the observed phenomenon. One reason can be arise from the morphology of deposited films. Nodule like structure with such a large voided boundaries and 100–200 nm grain size do not let charge carriers to drift. Charge carriers cannot pass the voided boundaries. They may be diffracted at the boundaries, which lead to higher resistivity of films. Another point is because of large amount of amorphous structure Cu3 N molecules in the samples as is shown by AFM micrograph. Usually amorphous materials have higher resistivity in comparison with crystalline materials.
10501-p3
The European Physical Journal Applied Physics
Fig. 5. The diagram of applied voltage versus current for measuring the resistance of sample 5.
Fig. 6. Electrical resistance of deposited films for different samples.
The other reason comes from the fact that the preferred orientation of lattice of deposited films belongs to N-rich planes. It is natural that resistance may be influenced by the content of nitrogen atoms in the film structure. More nitrogen leads to higher resistance. Change of resistance of samples is exactly similar with the variation of (100) planes and opposite with the variation of (111) planes in the structure of samples. According to XRD patterns for samples 1 to 3 the intensity of peaks belong to (100) planes decreases and the intensity of peak belong to (111) planes increases, leads to increasing the resistance. Sample 4 has the largest resistance because this film consists of (100) and (200) planes, which are both N-rich planes and there is not any (111) plane (Cu-rich) in the structure of this sample. Decreasing the resistance of sample 5 also can be explained similarly. These results confirm that the amount of N-rich planes in the lattice structure of samples is the most effective point in the resistance of generated films. 3.3 Band gap energy The optical energy gap Eg is another important quantity that characterizes semiconductors and dielectric materials since it has a paramount importance in the design and modeling of such materials [13]. The optical transmittances of Cu3 N films formed on glass substrate at various sputtering times are presented
Fig. 7. Transmittance of samples deposited on glass substrate.
in Figure 7. The optical transmittance of the films decreases with increasing the time of deposition which is the effect of increasing the film thickness. The wavelike patterns appear in the transmittance spectrum are due to the interaction of transmitted beams reflected from different interfaces i.e. film and glass substrate or film and air. All samples show low optical transmission or high absorption in UV and visible range up to about 550 nm, and show high optical transmission in near-infrared regions. The reflectance spectrum of samples is almost similar less than 0.05. From the spectral data of transmittance T and reflectance R, by using equation (1), the absorption coefficients α, is calculated for all samples. 1 (1 − R)2 (1 − R)4 2 α = ln + +R (1) d 2T 4T 2 in which d is the thickness of the deposited layer measured by profilometer [14]. Absorption coefficient of samples for different photon energy is illustrated in Figure 8a. It is clear that the absorption coefficient of samples sharply increases with increasing photon energy in the range of 1.8–2.4 eV. The optical energy gap was deduced from the intercept of the extrapolated liner part of the plot of (αE)1/2 versus the photon energy E with abscissa. This followed from the method of Tauc et al. [15] where αE = B(E − Eg )p .
(2)
In this equation α is the absorption coefficient, E is the photon energy, and B is a factor depends on the transition probability and can be assumed to be constant within the optical frequency range, and the index p is related to the distribution of the density of states. The index p has discrete values like 1/2, 3/2, 2 more depending on whether the transition is direct or indirect and allowed or forbidden, respectively. In the direct and allowed cases, the index p = 1/2 whereas for the direct but forbidden cases it is 3/2. But for the indirect and allowed cases p = 2 and for the forbidden cases it is 3 or more. Taking p = 2 corresponds to indirect allowed transitions in semiconductors, the band gap energy of deposited films are calculated using Tauc relation. Figure 8b displays plots of (αE)1/2 versus the photon energy E for Cu3 N films. A gradual formation of band gap energy and its increase can be seen
10501-p4
D. Dorranian and G. Mosayebian: Electrical properties of copper nitride thin films
(a)
(b)
Fig. 8. (a) The absorption coefficient and (b) the (αE)1/2 -E dependence of samples calculate the band gap energy.
in this figure. The magnitudes of the band gap energy of samples are between 1.76 to 1.91 eV which are presented in Table 1. Magnitudes of bandgap energy of different samples are in the range of obtained magnitudes for Cu3 N films in other reports. Variation of bandgap energy again can be due to the nitrogen contents in the lattice structure. By increasing the amount of nitrogen in the lattice structure or increasing the contribution of N-rich planes in the lattice structure of samples, the characteristics of samples tends from conductor to semiconductor leads to larger energy of band gap.
4 Conclusions The Cu3 N films were deposited on BK7 glass substrates by reactive direct current magnetron sputtering of a pure Cu target at room temperature but different time. Working gas was consisted of nitrogen and argon with equal contents. With this amount of nitrogen in working gas deposited films have a nodule like surface structure because of large number of amorphous Cu3 N molecules in the samples. X-ray diffraction measurements show that with increasing the sputtering time we have more (100) planes in the films orientation while the amount of (111) planes decreases. Obtained results in this experiment confirm that the resistivity and band gap energy of samples strongly depend on the content of N-rich and Cu-rich planes in the structure of deposited films. Following the results of reference [12], it can be observed that sputtering rate, film morphology and the bandgap energy of samples don’t change significantly with time of deposition while the nitrogen contents in working gas is 50%. It can be claimed
that the resistivity of films in this experimental situation can be just due to the preferred orientation of samples. Working gas with 50% nitrogen leads to more formation of (100) planes and increasing the resistivity of films.
References 1. M. Asano, K. Umeda, A. Tasaki, Jpn J. Appl. Phys. 29, 1985 (1990) 2. T. Maruyama, T. Morishita, Appl. Phys. Lett. 69, 890 (1996) 3. Z.G. Ji, Y.H. Zhang, Y. Yuan, C. Wang, Mater. Lett. 60, 3758 (2006) 4. G.H. Yue, P.X. Yan, J.Z. Liu, M.X. Wang, M. Li, X.M. Yuan, J. Appl. Phys. 98, 103506 (2005) 5. T. Maruyama, T. Morishita, J. Appl. Phys. 78, 4104 (1995) 6. Z.Q. Liu, W.J. Wang, T.M. Wang, S. Chao, S.K. Zheng, Thin Solid Films 325, 55 (1998) 7. L. Maya, J. Vac. Sci. Technol. A 11, 604 (1993) 8. G. Soto, J.A. Diaz, W. de la Cruz, Mater. Lett. 57, 4130 (2003) 9. J.F. Pierson, Vacuum 66, 59 (2002) 10. K.J. Kim, J.H. Kim, J.H. Kang, J. Cryst. Growth 222, 767 (2001) 11. Y. Du, A.L. Ji, L.B. Ma, Y.Q. Wang, Z.X. Cao, J. Cryst. Growth 280, 490 (2005) 12. D. Dorranian, L. Dejam, A.H. Sari, A. Hojabri, Eur. Phys. J. Appl. Phys. 50, 20503 (2010) 13. I.M. Odeh, J. Alloys Compd. 454, 102 (2008) 14. B.G. Streetman, Solid State Electronic Device, 3rd edn. (Prentice Hall Inc., New Jersey, 1990) 15. J. Tauc, R. Grigorovici, A. Vancu, Phys. Stat. Sol. 15, 627 (1996)
10501-p5
